This project uses structural equation modeling to examine the predictive validity of dentists' caries risk assessments and tests the congruence between patients' caries risk classifications and the caries preventive treatment they receive. The provision of caries preventive treatment linked to a patient's individually determined level of risk for the development of new lesions is increasingly being advocated as the standard of care for caries management. However, the predictive validity of dentists' assessment of caries risk, the degree to which patients with elevated risk classifications receive appropriate preventive treatment, and the effectiveness of preventive methods in patients with elevated risk have not been fully evaluated. These interrelated questions require a longitudinal multivariate approach. The proposed evaluation uses existing data from two staff model dental HMOs that have applied the risk-based management approach for several years to investigate the validity of caries risk assessment and the impact of subsequent treatment.